An energy-saving scheduling method supporting non-preemptive real-time task sets

Abstract

The invention discloses an energy-saving dispatching method capable of supporting a non-preemptive real-time task set. The method comprises the following steps: processing an initial task set, acquiring an initial deceleration parameter, acquiring candidate deceleration parameters, and distributing the deceleration parameters. On the premise of finishing non-preemptive real-time tasks before a deadline, by adopting the energy-saving dispatching method capable of supporting the non-preemptive real-time task set disclosed by the invention, the non-preemptive property of the task set is fully considered to obtain a relatively small deceleration parameter, thus enabling the task set to operate in a relatively low speed, achieving a better energy-conservation effect, effectively saving the energy consumption of an embedded system, and achieving relatively high non-preemptive task set real-time energy-conservation dispatching capability. Compared with an existing real-time system energy-conservation dispatching method, the method provided by the invention has the advantages that calculation load from high-priority tasks is accurately analyzed, one relatively small deceleration parameter is distributed to each task, and the energy consumption of the system can be obviously reduced.

Description

technical field [0001] The invention belongs to the technical field of real-time energy-saving scheduling, and more specifically relates to an energy-saving scheduling method supporting non-preemptive real-time task sets. Background technique [0002] Dynamic Voltage Scaling (DVS) is an effective low-power technology based on preemptive scheduling; however, due to hardware device or software configuration, in some cases, the cost of preemptive scheduling is too high. Large, for example, the load for preemptive operations is too large for the system (for example, in the commercial embedded real-time database system eXtremeDB, transaction scheduling is non-preemptive), or the system itself needs to support non-preemptive key segment tasks, you need to consider non-preemptive Energy-saving scheduling in the case of preemption; non-preemptive scheduling has the advantages of easy implementation, no synchronization overhead, and convenience for accurate analysis of task response ...

AI Technical Summary

Problems solved by technology

[0003] In the prior art, there is no energy-saving scheduling algorithm for completely non-preemptive task sets with static priorities. The two most relevant works are DS algorithm and USFI (UniformSlowdown algorithm with Frequency Inheritance), a dynamic voltage adjustment algorithm for synchronous tasks based on static priorities. , both DS and USFI algorithms are suitable for non-preemptive scheduling, but these two types of methods are based on the sufficient schedulability analysis of preemptible task sets considering task blocking to calculate task deceleration factors (normalized execution speed), and do not consider non-preemptive Due to the characteristics of preemptive scheduling, the deceleration factor is too conservative when calculating the deceleration factor, which cannot fully save the energy consumption of the system.

Images